New insights into the molecular mechanism of the action of diuretics.

The mechanisms of Na+, Cl− and 2 transport occur and can now explain previously poorly underalong the various nephron segments are depicted in stood genetic diseases. Figure 1. These schemes have all evolved from funcIn the proximal tubule inhibitors of carbonic anhydtional studies involving a large spectrum of methodolorase can be used to interfere with HCO3− absorption. gies [1,2]. According to these schemes the mechanisms Their effect is usually not very marked because the by which diuretics act in the kidney have been elaborensuing acidosis limits their inhibitory effect, and ated [3] and, in fact, in many instances the diuretics because more distal nephron segments, by increasing have been used to clarify the mechanisms depicted in the absorption at these sites, blunt the natriuresis and Figure 1. All diuretics inhibit the absorption of Na+, diuresis. Genetic defects in carbonic anhydrase II lead Cl− (HCO3−) and water by their interaction with to osteopetrosis connected with proximal-tubule-type luminal uptake mechanisms. Although it has been renal tubular acidosis [12]. Another possible target for shown that diuretics can also exert effects on tubule inhibition of proximal tubule absorption is the NHE-3 cells from the peritubule side at high concentrations, Na+/H+ exchanger (Figure 3A). Its sensitivity and these effects are probably not of pharmacological relevselectivity towards amiloride and ethylisopropylamilorance [4]. Diuretics are kidney specific, not because ide is rather poor [13]. However, very recently a new they interfere with proteins specific for the kidney, but derivative benzylguanidino amiloride has been shown because they are secreted into the lumen. This secretion to block this cotransporter with a very high affinity and the water absorption along the nephron contribute [14]. to the high free concentrations of diuretics in tubule The Na+2Cl−K+ cotransporter of the thick lumen. The mechanism whereby diuretics are secreted ascending limb is depicted schematically in Figure 3B and hence accumulated has been unravelled very [15,16 ]. This protein is closely related to the Na+Cl− recently [5–9] and is depicted in Figure 2. Diuretics cotransporter of the distal tubule [17]. However, it is are taken up across the basolateral membrane by an of clinical relevance that there exists no overlap in anion(e.g. furosemide, thiazides) or cation(amilorterms of pharmacology; loop diuretics bind only to the ide, triamterene) exchanger [5,6 ]. The exit step into Na+2Cl−K+ cotransporter but not to the Na+Cl− the lumen is not quite clear. It cannot be excluded that cotransporter, and thiazides bind only to latter but not this occurs via ATP-driven pumps such as those to the former. The Na+2Cl−K+ cotransporter exists described for the secretion of bile acids in the canalicuin two major forms. One is the colonic type and is lar membrane of the hepatocyte [10]. This would be found in almost every tissue. It participates in reguin agreement with old observations that a collapse in latory volume increase in almost every cell and is the secretion of anions is one of the first indicators of involved in the basolateral uptake of Cl− in exocrine proximal tubule damage, and it also explains why the glands. This type of cotransporter has an affinity response to, for example, loop diuretics is attenuated to diuretics (azosemide>bumetanide>furosemide) or abolished in renal failure [11]. Finally any other different from that present in the thick ascending limb organic anion accepted by the basolateral exchanger (bumetanide=torasemide>piretanide>furosemide= will compete with the secretion of diuretics [4]. azosemide) [4,18]. The other type of Na+2Cl−K+ Recently, the new methodologies of expression cloncotransporter appears to be specific for the kidney. ing have been used to identify the targets of diuretics Regulation of the Na+2Cl−K+ cotransporter is comat the molecular level. These efforts have been successplex. Current evidence suggests that it is upregulated, ful and have not only proven the correctness of schemes with activation of proteinkinase A, when cytosolic suggested previously, but have also revealed that genCl−-activity falls and cells shrink [19–22]. Whether etic defects in these transporters or related proteins all three signalling pathways converge or whether cell shrinkage is the common denominator is currently being studied. Correspondence and offprint requests to: Dr R. Greger, Mutations of this membrane protein can lead to Physiologisches Institut der Albert-Ludwigs-Universität, HermannHerder-Straße 7, D-79104 Freiburg, Germany. Bartter’s syndrome type I, resembling the same symp-